The incidence of ductal carcinoma in situ (DCIS) of the breast, a non-invasive form of breast cancer, has increased dramatically in the last 15 years. Its burden both on patients and on society has grown correspondingly. The optimal management of DCIS remains controversial because of the heterogeneity of the disease, the lack of randomized clinical trials comparing treatment strategies for women diagnosed with DCIS, the importance of patient preferences for possible outcomes and the uncertainty surrounding its natural history. Variations in the treatment of DCIS highlight the gaps in knowledge about the optimal management of the disease, gaps that have become increasingly important as the incidence of DCIS has increased. The cost implications of treatment variations also become substantial as DCIS is diagnosed more frequently. Ultimately, the variations in treatment result in differences in outcomes, including life expectancy, quality of life, and cost-effectiveness. We will examine the effects of various treatment strategies, including mastectomy with and without tamoxifen, and breast-conserving surgery with and without radiation and tamoxifen, on the following patient outcomes: DCIS recurrence rates, survival, costs, and quality of life. Decision analytic models will be used to estimate the cost-effectiveness and cost-utility of the various treatment strategies. Models will include patient preferences for DCIS and associated treatments obtained from primary data collection. Transition probabilities for the decision analytic models wilt be estimated from primary data using duration models and supplemented from the literature as necessary. Potential endogeneity in treatment selection will be corrected using instrumental variable techniques. The linked Medicare-SEER data will be used to examine the generalizability of our estimated transition probabilities. DCIS treatment costs will be estimated using Medicare data. Sensitivity analyses will be used to test the robustness of our models. The ultimate goal of our project is to identify the most cost-effective approaches to manage DCIS, taking into account a variety of clinical presentations and patient preferences, thus improving patient care and reducing the burden of the illness on society.